Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: ONT Re: Data Models, Ontologies, Logic




Jon and Matthew,

I, too, would like very much to get a handle on the notions you mention
below.  And I think the recent exercise in discussing "data model" has
helped a little bit. I agree with you both and with John Sowa
that the important thing is to be aware of how interested parties may
be using the term. And I also think that the purposes of why we
work with data, models, and theories ought to be kept in mind.

I remain curious about what useful distinctions can be drawn between
data models and ontologies, because (among other reasons) I think the 
question "Why bother with an ontology when we can use a data model instead?" 
deserves to be answered.

I wonder if it is worthwhile to consider the differences between 
ontologies that are created to understand the world but that aren't 
meant to be computed with, ontologies that are created to reflect 
some part of the world and _are_ meant to be computed with, and data models.
I am still reading what John Sowa wrote the other day about Matthew's
distinctions...

I appreciate very much the discussion this thread.

Jim

-------------------------------------------------------------------------

JA - Jon Awbrey


JA: > The problem, as I see it, is to reconnect the different senses
    > of words like "data", "description", "model", "theory", "world"
    > that are employed by all brands of viewers of the current scene.
    > 
    > John Sowa has pretty picture of this that keeps coming back to me:
    > http://www.jfsowa.com/figs/mthworld.gif


JA: > But I feel like we may be getting distracted with this quibble over 
    "model".
    > I'm already used to the way that it's used in six or seven different 
    fields.
    > And I'm the one who said that ambiguity would be unavoidable, if you 
    recall.
    > So that is not the problem, not for me.  The problem is keeping clear 
    about
    > what senses of the word we are using at any given time, and not 
    letting any
    > one usage prevent us from remembering the purpose of data, models, 
    theories.

-----------------------------------------------------------------------------